You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using the library with a custom Bonsai workflow (repository is here, but let me know i you have any questions), we have encountered a thread deadlock after running long acquisition sessions (>4hours). From the point of view of bonsai, the lock manifest itself as something that stops the update of all visualizers and prevents the "stop" button from being pressed.
After making a quick memory dump from the process while in its frozen state, I realized that one of the threads that was locked was the one used by the PolledBNO055Data operator. (I am happy to provide the mem dump if you want, but its quite a large file (2gb).)
In the meantime, I tried a solution that seems to have fixed it (or at least no crashes so far, including a full 12hour, 0 twisted turns on the wire, session :P)
), I noticed that the 0-argument overload is instantiating a NewScheduler. As I have had trouble with these before, I instead used the second overload with a Timer (which by default uses the DefaultScheduler, if I am not mistaken). This change has been sufficient to stop the deadlocks so far.
Hope it helps!
The text was updated successfully, but these errors were encountered:
Heya!
Following a chat with @bparks13...
When using the library with a custom Bonsai workflow (repository is here, but let me know i you have any questions), we have encountered a thread deadlock after running long acquisition sessions (>4hours). From the point of view of bonsai, the lock manifest itself as something that stops the update of all visualizers and prevents the "stop" button from being pressed.
After making a quick memory dump from the process while in its frozen state, I realized that one of the threads that was locked was the one used by the PolledBNO055Data operator. (I am happy to provide the mem dump if you want, but its quite a large file (2gb).)
In the meantime, I tried a solution that seems to have fixed it (or at least no crashes so far, including a full 12hour, 0 twisted turns on the wire, session :P)
Going through the source of the operator (
bonsai-onix1/OpenEphys.Onix1/PolledBno055Data.cs
Line 57 in 3c6702b
Hope it helps!
The text was updated successfully, but these errors were encountered: